This a randomized, double-blind, sham-controlled trial to determine the effectiveness of celiac plexus neurolysis (CRN) as an adjunct to standard medical therapy for pain control in patients with unresectable pancreatic cancer and moderate to severe pain. Nearly 100% of patients with pancreatic cancer experience pain. Despite the availability of narcotic analgesics, up to 50% of patients have poorly controlled pain. A 1993 randomized placebo-controlled trial of surgical CRN, performed at the time of exploratory laparotomy, demonstrated CRN to be safe and effective. Recent advances in imaging methods (CT, MRI, EUS) have allowed better pre-operative identification of patients with unresectable disease, such that patients rarely require exploratory laparotomy. For this reason a minimally invasive method of pain control in pancreatic cancer which could be performed in conjunction with staging and tissue confirmation, would be a major clinical advance. Endoscopic ultrasound is a widely used, minimally invasive, highly accurate and safe method of staging, and tissue biopsy of pancreatic cancer. Phase II, uncontrolled trials suggest that EUS is also capable of delivering effective celiac plexus neurolysis (EUS-CPN). To date, there are no published controlled trials or EUS-CPN. This study will enroll patients who are undergoing diagnostic EUS-FNA procedures for staging of pancreatic cancer. Those found to be unresectable would be randomized "intra-op" to either sham or real EUS-CPN. Outcomes will include pain scores, quality of life, narcotic use and survival. The primary outcome will be a reduction in pain at 1 month. Secondary outcomes will include the durability of pain control, quality of life, narcotic use and narcotic side effects. [unreadable] [unreadable]